The International Journal of Psychosocial Rehabilitation
Factors Predicting Overall Life Satisfaction for People
with Long-Term Mental Illness

  Citation:
Young , K.M.   (2004).  Factors Predicting Overall Life Satisfaction for People with Long-Term
 Mental Illness Factors.    International Journal of Psychosocial Rehabilitation. 9, 23-35.

E-mail: Youngkimwandaniel@hotmail.com

---

Abstract
This study attempts to explore those factors and program elements leading to better overall life satisfaction for people with long-term mental illness. The research sample included 88 patients, coming from mental hospitals and a large residential home in Hong Kong. Quality of Life Interview, Perceived Social Support and Sense of Freedom were modified, developed and adopted for data collection purpose. Most of the studied sample suffered from schizophrenia and had been ill for more than twenty years. Results showed that three factors including: number of hospitalization, perceived sense of freedom and social support could explain about one third of the variance in overall life satisfaction of research sample.   Based on these findings, a quality of care model, which emphasis on preventing relapse & hospitalization, enhancing sense of freedom and strengthening social support, has been proposed.  This quality of care model is suitable for residential home care setting which aims at promoting the quality of life for people with long-term mental illness.
 
Key words: Quality of Life, Mental Illness

---

Introduction
Few studies have been done in identifying those program elements predicting subjective life satisfaction for people with long-term mental illness. Mental health research is often confronted with the problem that the intervention or program under study takes place in a “black box”, which makes it difficult to assess the relative importance of various components of the intervention [1].  Moreover, even with programme that have demonstrated successful outcomes, it is difficult to specify which components of the intervention are responsible for the observed outcomes [2]. This study attempts to explore those factors and program elements leading to better overall life satisfaction (below referred as Overall QoL) for people with long-term mental illness.  
 
Research Design and Sample
People with long-term mental illness, who were living in two traditional mental hospitals and a large residential home in Hong Kong, were invited to join in this research project. Finally, 88 people with long term mental illness had given their written consent to this research project and had been successfully interviewed by the author, including: Hospital group (n=43), and discharged group (n=45). The Hospital group was those who were ready for discharge and interviewed by researchers when they were living at two large traditional mental hospitals. Discharged group was those who had lived in a large residential home for about one year. That large residential home has been opened on 1-4-1996.  It can accommodate two hundreds people with long-term mental illness. Various kinds of professional and non-professional staff have been employed to provide 24-hour personal care services for its residents. Staff to resident ratio is about 1:3.  The QoL data of the research sample was collected during between April 1996and December 1997.
 
Data Collection
Several measuring scales have been used in this study. 
A)  Quantitative Measuring Instruments Used

1)  Quality of Life Interview (Modified)
Quality of Life Interview [3] is adopted and modified in this study.  The reliability and validity of the scale are well established [4] and the scale has been widely used [5-9]. There are two modifications. The first modification is the addition of two domains, i.e. staff and vocational training, with their related objective and subjective indicators.  The second modification is the modification of the objective indicators of living environment so as to suit the special living environment in mental hospital and residential home settings in Hong Kong. The modified scale has 56 items, and contains objective QoL indicators, subjective QoL indicators as well as overall life satisfaction.
 
2)        Perceived Social Support Scale
The multi-dimensional scale of perceived social support [10] is adopted and modified in this study. The validity and reliability of the scale has been tested to be satisfactory [11]. In this study, respondent is asked whether family, friends and staff would provide assistance in seven occasions which are adapted to hospital and residential care setting.
 
3)        Sense of Freedom
The author has developed this measuring scale for the use in hospital and residential setting.  It divides into two main sub-scales: perceived sense of freedom and perceived restriction.  The perceived restriction sub-scale is a binary yes / no scale.  The perceived sense of freedom sub-scale covers nine areas: money management, going out, bed time, bathing time, dressing, hair cutting, meal and vocational training as well as overall freedom. Each of these items is measured in a four-point scale.
 
4)  Social Behaviour Schedule (SBS)
The Social Behaviour Schedule (SBS) is chosen for use in this study as this scale is especially designed for measuring the functioning in 21 areas such as hygiene, initiating conversations, depression, violence, etc. for long-stay psychiatric patients [12]. The reliability and validity of the scale has been reported to be satisfactory.
 
5)  Specific Level of Functioning Assessment (SLOF)
The Specific Level of Functioning Assessment (SLOF) is designed to measure more directly observable behavioural functioning and daily living skills of people with chronic mental illness [13]. The reliability and validity of the scale has been reported to be satisfactory.

 
 
Reliabilities of the Measuring Scales
The reliabilities of self-developed scales, including: Quality of Life Interview (Modified), Perceived Social Support, Perceived Sense of Freedom and their respective  sub-scales are measured by means of internal consistency using Cronbach’s alpha.  Table 1& 2 gives the results.
 
As shown in the Table 1 & 2,  the internal consistencies of these scales and their subscales have reached an acceptable level of reliability for statistical analyses purpose. In addition, the internal consistencies of Quality of Life Questionnaire and Quality of Life Interview [12] are comparable.

Table 1 Comparison of Cronbach’s alpha ( a ) of Quality of Life Interview (Modified) and Lehman’s Quality of Life Interview

	Cronbach’s alpha ( a ) of Quality of Life Questionnaire (current study)	Cronbach’s alpha ( a ) of Quality of Life Interview (Lehman, 1983)
Subjective QoL Indicators
Living Situation	.80	.83
Daily Activities	.89	.81
Family Relationship	.65	.89
Social Relationship	.86	.80
Staff	.82	N.A.
Finance Situation	.93	.87
Work / Vocational Training	.79	N.A.
Legal and Safety Issue	.65	N.A.
Health	.74	.83
Objective QoL Indicators
Daily Activities	.55	.67
Family Relationship	.74	N.A.
Social Relationship	.68	.76
Finance Situation)	.69	.81
Overall Life Satisfaction
Overall Life Satisfaction	.62	N.A.

 

Table 2 Cronbach’s alpha ( a ) of Perceived Social Support and Sense of Freedom

Measuring Scale	No. of Item	Internal consistency (Cronbach’s alpha )
Perceived Social Support
Perceived Social Support Scale (Whole Scale)	21	.92
Perceived Family Support Sub-scale	7	.94
Perceived Friend Support Sub-scale	7	.87
Perceived Staff Support Sub-scale	7	.80
Sense of Freedom
Sense of Freedom (Whole Scale)	18	.77
Perceived Restrictiveness (Sub-scale)	9	.61
Perceived Sense of Freedom (Sub-Scale)	9	.66

 

Construct Validity of the Quality of Life Interview (Modified) scale
There is a consistent finding in literature that subjective and objective QoL  measures are only weakly inter-related in the same life domain [14]. This study would like to test the construct validity of the scale by exploring: whether or not the subjective and objective quality of life data obtaining by the Quality of Life Interview (Modified) scale, are inter-related in the same domain.
 
Results showed that the subjective QoL and objective QoL are not related in the domains of daily activities and social relationship, family and social relationship, while they are weakly or moderated related in the domain of staff relationship (r =-.208, p=.028), and finance (r =.429; p=.000) respectively. This result shows that the Quality of life Interview (Modified) scale pass the test of construct validity and it can be taken as a valid measure.
 
Characteristics of the research sample
Below describes the characteristics of the studied group.  Results showed that the hospital and discharged groups did not differ in almost all areas of life, including demographic factors, psychiatric impairment and social functioning.  They differed in their sex ratio and physical functioning only.  The hospital group was predominately male while discharged group was predominated female. Also the hospital group was found to have poorer ability in their use of legs and hands as compared with that of discharged group. Thus the characteristics of these two studied groups were more or less the same.
 
Taking together, the studied groups were predominately male, schizophrenic and single.  All of them were under 65 years of age (mean  47-48 years).  Most of them had reached primary school level or above. In term of social handicap, most were single or divorced, and lack of social  support had few contacts with their families. Also most of them lack of employable skills and had to rely on governmental social security schemes to support their lives.
 
In term of psychiatric impairment, almost all of them suffered from schizophrenia.  Their illnesses were so chronic that most of them had been ill for more than 20 years, and all of them had a current continuous hospital stay for more than one year. Over one fifth showed at least one significant behavioural problem, including both positive and negative symptoms, that needed staff’s intervention, as measured by SBS.  In terms of social functioning, over half of them showed at least three problems in their functioning, especially in work skills, community living skills, and interpersonal relationships, as measured by SLOF.
 
 
Research Results
 
Correlation Analysis
A.     Demographic and Medical variables
 
B.  Current Symptomatology & Functioning
 
C.     Objective & Subjective QoL Indicators
 
D. Social Support & Sense of Freedom
   
Factors leading to a better Subjective Overall Life Satisfaction - Regression Analyses
A  variety of regression models can be constructed from the same set of variables.  In selecting those variables entering into the regression analysis, theoretical consideration plays a very important role. Regression models are specified in SPSS for Windows using forward selection procedures. Two regression models are constructed.
 
First Regression Equation
The first model constructed examines the best predictors of subjective overall life satisfaction for the studied sample.  Three types of variables are selected and entered into the first regression equation.  The first group is the  clinical variables, i.e. number of hospitalization.  The second group involves the  objective QoL indicators, mainly objective daily activities and functioning.  The final group involves  subjective QoL indicators, including:  satisfaction with daily activities and functioning, satisfaction with safety at current living place.
 
Then by using hierarchical analyses, we can explore the relative importance of different category of variables in the regression equation.  In doing this, clinical variables i.e. number of hospitalizations, is first entered into the equation, followed by the  objective QoL indicators ( objective daily activities and functioning) and finally the  subjective QoL indicators.  The regression analysis is then done step by step for the above categories of variable. 
 
The results show that these variables altogether could explain about half (46.7%) of variance of subjective overall life satisfaction. Also it shows that the clinical variables accounted for  18.4 % of variance in  Overall QoL. Objective daily activities and functioning explains an additional 8.2% of the variance,  and the  subjective QoL indicators explain an additional 20% of the variance. (Please refer to Table 2 & 3 for reference).  Thus the individual subjective QoL indicators are the strongest variables in explaining the variance of Overall QoL.   This result is consistence with other research studies which have found that subjective QoL indicator is the strongest predictor of Overall QoL  [15-17].

Table 3 Model Summary of the hierarchical regression equation

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	0.441	0.195	0.185	1.2589
2	0.534	0.285	0.267	1.1934
3	0.702	0.493	0.467	1.018

a  Predictors: (Constant), Number of hospitalization
b  Predictors: (Constant), Number of hospitalization, Objective QoL in daily activities & functioning
c  Predictors: (Constant), Number of hospitalization, Objective QoL in daily activities & functioning,
                             Subjective QoL in daily activities & functioning, Subjective QoL in personal security.

Table 4 ANOVA of the hierarchical regression equation

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	31.039	1	31.039	19.584	0
	Residual	128.377	81	1.585
	Total	159.416	82
2	Regression	45.486	2	22.743	15.97	0
	Residual	113.929	80	1.424
	Total	159.416	82
3	Regression	78.582	4	19.645	18.957	0
	Residual	80.834	78	1.036
	Total	159.416	82

a  Predictors: (Constant), Number of hospitalization
b  Predictors: (Constant), Number of hospitalization, Objective QoL in daily activities & functioning
c  Predictors: (Constant), Number of hospitalization, Objective QoL in daily activities & functioning, Subjective QoL in
                             daily activities & functioning, Subjective QoL in personal security, Subjective satisfaction with freedom
d  Dependent Variable: Overall QoL

Second Regression Equation
A key interest in this research is in exploring those factors, which could be modified by staff,  in producing better subjective overall life satisfaction for people with long-term mental illness. In order to do this it is necessary to re-analyze the data omitting the individual subjective QoL indicators  as independent variables in the regression equation.
 
Three variables are selected for entry by the forward selection procedure.  These variables are: number of hospitalizations, perceived staff support and overall sense of mastery.  This second regression equation explains about one third (30.2%) of the variance in Overall QoL, which  shows that predictive power of this second regression equation is acceptable (Please refer to Table 5 to Table 7).
 
If predicting Overall QoL individually,  these three variables also yield satisfactory predictive power (please refer to Table 9).  Number of hospitalization has the strongest predictive power and could explain about one fifth (18.5%) of variance in Overall QoL, while sense of freedom explains 10.4% and perceived staff support explains 9.4% respectively.

Regression Analysis 2

 Table 5 Model Summary of the 2^nd regression equation

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	0.445	0.198	0.188	1.2641
2	0.539	0.291	0.273	1.1963
3	0.572	0.327	0.302	1.1723

a  Predictors: (Constant), number of hospitalization
b  Predictors: (Constant), number of hospitalization, perceived staff support
c  Predictors: (Constant), number of hospitalization, perceived staff support, sense of freedom

Table 6 ANOVA of the 2^nd regression equation

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	31.543	1	31.543	19.739	0
	Residual	127.836	80	1.598
	Total	159.378	81
2	Regression	46.326	2	23.163	16.186	0
	Residual	113.052	79	1.431
	Total	159.378	81
3	Regression	52.189	3	17.396	12.659	0
	Residual	107.189	78	1.374
	Total	159.378	81

a  Predictors: (Constant), number of hospitalization
b  Predictors: (Constant), number of hospitalization, perceived staff support
c  Predictors: (Constant), number of hospitalization, perceived staff support, sense of freedom
d  Dependent Variable: overall life satisfaction

Table 7  Predictive Power of individual variables of: Staff Support, Sense of freedom, Number of hospitalization

	Variable	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	Sense of freedom	0.306	0.115	0.104	1.3187
2	Staff Support	0.323	0.104	0.094	1.3036
3	Number of hospitalization	0.441	0.195	0.185	1.2589

a  Predictors: Staff Support, Sense of freedom, Number of hospitalization
b  Dependent Variable: overall life satisfaction

In addition, sense of freedom and perceived staff support are found to have mediating effects on the relationships between patients’ subjective overall life satisfaction with objective and subjective quality of life in living situation.  As mentioned above, patients’ objective and subjective quality of life in living situation are found to have a significant and moderate correlation with subjective overall life satisfaction (Table 8).  However, after controlling sense of freedom and perceived staff support, these correlations turn to be non-significant (Table 8).  It seems that when patients evaluating their living situation, they would use these two factors i.e. sense of freedom and perceived staff support as two important criteria in evaluating their objective and subjective quality of life in this area.

Table 8 Correlations of Overall subjective life satisfaction with Objective and Subjective QoL in living situation

		Without any controlling variables		Controlling for Sense of freedom and Perceived staff support
Item	Test	Value	Significance	Value	Significance
Objective QoL in living situation	Pearson Coefficient	-0.223	.037**	-0.1356	0.216

**  Correlation is significant at the 0.05 level (2-tailed).

Discussion

The first regression model explains about half of variance of subjective overall life satisfaction and finds that the individual subjective QoL indicators are the strongest variables in explaining the variance of Overall QoL.   These results are consistence with other research studies which have found that subjective QoL indicator is the strongest predictor of Overall QoL  [15-17]. This finding adds evidence to the validity of the modified Quality of Life Interview used in this study.
 
The second regression model consists of three variables, including: number of hospitalizations, perceived staff support and overall sense of mastery, and can explain about one third of the variance in Overall QoL.  This finding is very helpful in identifying those program elements leading to a better Overall QoL for people with long-term mental illness.
 
A Quality of Care Model for residential home
Based on the findings of the second regression model, a quality of care model is proposed to include those program elements as well as intervention methods leading to a better Overall QoL for people with long-term mental illness. This quality of care model is particular useful for residential services. Figure 1 gives a graphical presentation of this model.
 
There are three core programme elements in this model.  These three core elements are: preventing relapse and hospitalization,  strengthening social support and  enhancing sense of freedom. For each core program element, there are various kinds of related intervention methods. For example, in preventing relapse and hospitalization, medication, social skill training, cognitive behavioural therapy and family psycho-education are well known effective intervention methods  [18-22]. For detail discussion, please refer to the unpublished thesis written by the author  [23].


Figure 1. A Quality of Care Model for long-stay psychiatric patients.        

Quality of Care                 Quality of Life

Reflection on methodology

The study of subjective quality of life is largely based on the life satisfaction model as proposed by Andrew & Withey [24] as well as Campbell [25] and then adopted by Lehman [26] in the study of quality of life for people with mental illness.  According to this model, quality of life is a cognitive experience in which an individual compares his perception of his present situation to a situation which he aspires to or expects. WHOQOL group pointed out that quality of life is both subjective and sensitive to the cultural and value systems where they live [27]. In exploring the predicators of subjective quality of life, it is difficult to identify suitable predicators as different respondents may have totally different value systems which affected their expectations, goals and standards in relation to their quality of life.
 
In mental health field, few studies have been done in identifying program elements predicting overall life satisfaction for people with mental illness.  Levitt [16] was the first study to explore this issue.  In his study, subjective satisfaction with adult education of rehabilitation service was found to predict overall life satisfaction for people with mental illness receiving adult education. However, in that study, the reasons why respondents were satisfied with adult education program remained unanswered.  Rosenfield [28] successfully identified specific program elements in predicting overall life satisfaction for people with mental illness attending a clubhouse.  It was found that vocational rehabilitation, financial support services and an empowerment approach to psychiatric treatment were good predictors.
 
As shown in this study, three key program elements, including: preventing relapse and hospitalization, promoting social support and enhancing sense of freedom are found to be good predicators of overall life satisfaction for people with long term mental illness. The identification of these key program elements with  the related quality of care model could then be generalized to other residential homes.

In this study, it is shown that it would be easier for study in exploring predicators of quality of life to include a more or less homogeneous studied sample receiving in a similar kind of supportive service. As shown above, the characteristics of both studied groups, i.e. hospital and discharged group were very similar. Also, they lived in the same residential home or mental hospitals with similar characteristics. In such case, each individual respondent of a homogeneous studied group may share similar value system with each other and has similar expectations and common concerns about his quality of life. So, the identification of predicators of quality of life for such a homogeneous group would be easier and feasible. This kind of research strategy seems promising in identifying predicators of quality of life. Thus more research using similar kind of research strategy are needed in this area so as to find out the ways to further promote the quality of life for people with mental illness supported by different community care programs.

---

1. Mechanic D. (1996) Emerging issues in international mental health services research Psychiatric Service 47: 371-375

2. Brekke J.S. (1988) What do we really know about community support programs? Strategies for better monitoring Hospital and Community Psychiatry 39: 946-952

3. Lehman A.F. (1983) The effects of psychiatric symptoms on quality of life assessment among the chronic mentally ill Evaluation and Program  Planning 6:143-151

4. Lehman A.F. (1988) a quality of life interview for the chronically mentally ill Evaluation and Program Planning 11:51-62

5. Barry M.M. Crosby C. (1996) Quality of life as an evaluative measure in assessing the impact of community care on people with long term psychiatric disorders British Journal of Psychiatry 168: 210-216

6. Sullivan G., Wells K., Lecke B. (1991) Quality of life of seriously mentally ill persons in Mississippi Hospital and Community Psychiatry 42:752-754

7. Levitt A.J., Hogan T.P., Bucosky C.M. (1990) Quality of life in clinically mentally ill patient in day treatment Psychological Medicine 20:703-710

8. Simpson C.J., Hyde C., Farragher E.B. (1989) The chronically mentally ill in community facilities British Journal of Psychiatry 151:77-82

9. Warner R., Huxley P. (1993) Psychopathology and quality of life among mentally ill patient in the community British Journal of Psychiatry 163:p.505-509

10. Zimet G.D., Dahlem N.W., Zimet S.G.H., et al (1988) The multi-dimensional scale of perceived social support Journal of personality assessment 52: 30-41

11. Kazarian S.S., McCabe C.B. (1991) Dimensions of social support in MSPSS Journal of Community Psychology 19: 150-160

12. Wykes T., Stuat E. (1986) The measurement of social behaviour in psychiatric patients: an assessment of the reliability and validity of SBS schedule British Journal of Psychiatry 148: 1-11

13. Schenider L.C., Struening E.L. (1983) SLOF: a behavioural rating scale for assessing the mentally ill Social Work Research and Abstract 13: 9-21

14. Barry M.M., Crosby C., Mitchell D. (1992) Quality of life issues in the evaluation of mental health services In Trent D.R.(ed) Promotion of mental health vol 1 Avebury

15. Lehman A.F. (1988) a quality of life interview for the chronically mentally ill Evaluation and Program Planning 11:51-62

16. Levitt A.J., Hogan T.P., Bucosky C.M. (1990) Quality of life in clinically mentally ill patient in day treatment Psychological Medicine 20:703-710

17. Oliver J. (1992) The social directives care: development of a quality of life profile for use in community services for mentally ill Social Work and Social Science Review 3:5-45

18. Anthony W.A., Liberman, P.R. (1990) Principles and practice of psychiatric rehabilitation in Liberman R.P. Handbook of psychiatric rehabilitation New York Pergamon

19. Benton M.K., Schroeder H.E. (1990) Social Skill training with schizophrenics: a meta analytic evaluation Journal of Consulting and Clinical Psychology 58: 741-747

20. Dilk M.N., Bond G.R. (1996) Meta-analytic evaluation of skills training research for individuals with severe mental illness Journal of Consulting and Clinical Psychology 64: 1337-1346

21. Heinssen R.K., Liberman R.P., Kopelowicz A. (2000) Psychosocial skills training for schizophrenia: lesions from the laboratory Schizophrenia Bulletin 26: 21-46

22. Dickerson F.B. (2000) Cognitive behavioural psychotherapy for schizophrenia: a review of recent empirical studies Schizophrenia research 43: 71-90

23. K.W.Young (2003) Quality of Life of Discharged Long-Stay Psychiatric Patients PhD Thesis

24. Andrews F.M., Withey S.B. (1976) Social indicators of well being: America’s perception of life quality New York Plenum Press

25. Campbell A., Converse P.E., Rodgers W.L. (1976) The quality of American life New York: Rissell sage Foundation

26. Lehman A.F. (1988) a quality of life interview for the chronically mentally ill Evaluation and Program Planning 11:51-62

27. WHOQOL group (1995) The world health organization quality of life assessment (WHOQOL): position paper from the world health organization Social Science and Medicine 41: 1403-1409

28. Rosenfield S. (1992) Factors contributing to subjective quality of life of chronically mentally ill Journal of Health and Social Behaviour 33:299-315

---